INDEX

Explanations

names of actors and directors

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals involved in film or media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

nings

-0.70

tions

-0.69

 newsletters

-0.66

lies

-0.65

BALL

-0.65

leaders

-0.64

coins

-0.63

ramid

-0.62

bleacher

-0.62

 ballpark

-0.62

POSITIVE LOGITS

 Films

0.83

 Runtime

0.82

 Actress

0.80

 actress

0.80

 actor

0.79

 Ridley

0.76

 Spielberg

0.76

 Winc

0.75

Clo

0.74

 Whedon

0.74

Activations Density 0.140%

names of actors and directors

proper nouns, particularly names of individuals involved in film or media

No Comments

No Known Activations