INDEX

Explanations

instances of the word "stated" or its variants

oai_token-act-pair · gpt-3.5-turbo

phrases involving declarations or formal statements

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

Enlarge

-0.66

vernment

-0.65

erc

-0.64

fen

-0.60

lio

-0.60

ective

-0.59

icons

-0.59

edge

-0.58

Battery

-0.58

jar

-0.58

POSITIVE LOGITS

 unequivocally

1.39

 emphatically

1.18

 bluntly

1.15

 plainly

1.14

 categor

1.09

 explicitly

0.96

 unamb

0.93

 confidently

0.92

 definitively

0.91

 clearly

0.91

Activations Density 0.075%

instances of the word "stated" or its variants

phrases involving declarations or formal statements

No Comments

No Known Activations