INDEX

Explanations

phrases related to specific technical or specialized terms or concepts

oai_token-act-pair · gpt-3.5-turbo

terms associated with various cultural references and topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 grate

-0.56

endum

-0.56

 Mehran

-0.55

rower

-0.55

 summed

-0.54

 welcomed

-0.53

 testified

-0.53

 saddened

-0.53

undrum

-0.52

 reacted

-0.52

POSITIVE LOGITS

clips

0.63

$.

0.60

cycles

0.60

().

0.59

abs

0.59

ds

0.57

thood

0.56

gage

0.56

max

0.55

gress

0.55

Activations Density 1.055%

phrases related to specific technical or specialized terms or concepts

terms associated with various cultural references and topics

No Comments

No Known Activations