INDEX

Explanations

mentions of parts, shares, or quantities of something

oai_token-act-pair · gpt-3.5-turbo

references to fractions or segments of items or concepts

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

 Fighters

-0.71

Trend

-0.68

generic

-0.68

verbs

-0.67

ATHER

-0.65

raid

-0.63

lear

-0.61

æ©

-0.59

Iran

-0.59

friends

-0.59

POSITIVE LOGITS

 thereof

1.05

meal

0.79

 portions

0.77

edly

0.76

ials

0.76

 ILCS

0.73

itol

0.73

xual

0.72

icularly

0.71

xit

0.70

Activations Density 0.013%

mentions of parts, shares, or quantities of something

references to fractions or segments of items or concepts

No Comments

No Known Activations