INDEX

Explanations

phrases or sentences that contain specific names, potentially related to articles, news reports, or discussions

oai_token-act-pair · gpt-3.5-turbo

references to a specific object or term likely associated with titles or names in a structured format

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Leilan

-0.76

igil

-0.75

arium

-0.72

enery

-0.71

Magikarp

-0.70

forward

-0.69

 Sons

-0.69

forth

-0.69

burg

-0.66

ivities

-0.66

POSITIVE LOGITS

JECT

1.32

BY

1.06

OB

1.02

ject

0.97

tained

0.96

rien

0.91

ooth

0.91

lique

0.91

ilib

0.90

acter

0.88

Activations Density 0.007%

phrases or sentences that contain specific names, potentially related to articles, news reports, or discussions

references to a specific object or term likely associated with titles or names in a structured format

No Comments

No Known Activations