INDEX

Explanations

phrases indicating statements of truth

oai_token-act-pair · gpt-3.5-turbo

statements asserting the truth of various claims

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rador

-0.77

hens

-0.76

Pages

-0.76

uled

-0.75

aida

-0.74

adish

-0.72

iazep

-0.71

edin

-0.69

onut

-0.69

acent

-0.67

POSITIVE LOGITS

 believers

0.88

 believer

0.83

terday

0.80

hood

0.76

 regardless

0.73

 Nationwide

0.68

reading

0.67

 everywhere

0.67

 universally

0.65

 Detective

0.65

Activations Density 0.024%

phrases indicating statements of truth

statements asserting the truth of various claims

No Comments

No Known Activations