INDEX

Explanations

parts of names or terms

oai_token-act-pair · gpt-3.5-turbo

words related to car parts or automotive terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 indebted

-0.64

prime

-0.57

krit

-0.56

 Blumenthal

-0.56

factor

-0.56

 Imran

-0.54

cue

-0.53

 Revenge

-0.53

 animate

-0.53

ional

-0.53

POSITIVE LOGITS

utical

0.85

onne

0.84

heid

0.77

bean

0.76

ãĤ¼ãĤ¦ãĤ¹

0.74

ohyd

0.74

steen

0.72

imeter

0.72

anwhile

0.68

reau

0.68

Activations Density 0.191%

parts of names or terms

words related to car parts or automotive terms

No Comments

No Known Activations