INDEX

Explanations

proper nouns or initials related to organizations or projects

oai_token-act-pair · gpt-3.5-turbo

references to specific organizations or research firms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Eisen

-0.77

flies

-0.76

lust

-0.72

 Apprentice

-0.71

thumbnails

-0.65

 Cecil

-0.64

âĸĳâĸĳ

-0.64

athan

-0.64

enegger

-0.64

bell

-0.64

POSITIVE LOGITS

KP

1.05

rint

0.96

_.

0.92

PK

0.88

dh

0.85

PK

0.82

rompt

0.81

olicy

0.79

©¶æ

0.79

srf

0.76

Activations Density 0.010%

proper nouns or initials related to organizations or projects

references to specific organizations or research firms

No Comments

No Known Activations