INDEX

Explanations

adjectives and nouns describing the quality or intensity of various situations or attributes

oai_token-act-pair · gpt-3.5-turbo

phrases that express states of being or evaluations

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

zar

-0.84

cffffcc

-0.78

BUS

-0.73

 CrossRef

-0.72

pter

-0.72

vier

-0.71

tyard

-0.71

 Calm

-0.71

yna

-0.70

isky

-0.69

POSITIVE LOGITS

 warranted

0.67

 delusion

0.65

 respective

0.64

 entitlement

0.63

 fian

0.61

 Sandwich

0.61

 ever

0.60

 individual

0.60

 pill

0.60

 depend

0.59

Activations Density 0.116%

adjectives and nouns describing the quality or intensity of various situations or attributes

phrases that express states of being or evaluations

No Comments

No Known Activations