INDEX

Explanations

key terms related to providing evidence or demonstrating truth

oai_token-act-pair · gpt-3.5-turbo

instances of the word "proof" in various contexts

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

 contrace

-0.73

 livest

-0.64

 newsletters

-0.63

ecycle

-0.61

cffffcc

-0.59

 Peninsula

-0.58

 incent

-0.57

leton

-0.56

ideshow

-0.54

JO

-0.54

POSITIVE LOGITS

reading

1.51

reader

1.30

read

1.27

reads

0.93

positive

0.82

 thereof

0.80

proof

0.76

alling

0.74

itious

0.74

 positive

0.73

Activations Density 0.028%

key terms related to providing evidence or demonstrating truth

instances of the word "proof" in various contexts

No Comments

No Known Activations