INDEX

Explanations

the word "True" in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "true"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

uled

-0.79

acco

-0.77

Pages

-0.77

ONES

-0.76

adish

-0.75

ocene

-0.75

Corp

-0.74

served

-0.73

chains

-0.70

oleon

-0.70

POSITIVE LOGITS

 believers

0.93

 believer

0.90

stic

0.76

ignment

0.73

ll

0.72

 positives

0.71

çĭ

0.70

polit

0.70

 freshman

0.69

sell

0.69

Activations Density 0.017%

the word "True" in various contexts

instances of the word "true"

No Comments

No Known Activations