INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to actions and interactions involving objects and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

je

-0.82

ihu

-0.76

jar

-0.73

Font

-0.72

jan

-0.72

åĬ

-0.71

æī

-0.71

Tar

-0.71

cript

-0.70

kar

-0.69

POSITIVE LOGITS

anwhile

0.98

 alot

0.86

ALL

0.82

NOW

0.75

 entirety

0.74

all

0.73

etheless

0.70

 totality

0.70

 moreover

0.70

atter

0.70

Activations Density 0.250%

proper nouns

references to actions and interactions involving objects and entities

No Comments

No Known Activations