INDEX

Explanations

adjectives and phrases expressing opinions or evaluations

oai_token-act-pair · gpt-3.5-turbo

statements expressing serious concerns or issues

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

bats

-0.54

alks

-0.48

testers

-0.46

 Languages

-0.46

keys

-0.46

 slides

-0.45

 bowls

-0.45

anas

-0.45

 Talks

-0.44

 keyboards

-0.44

POSITIVE LOGITS

soType

0.65

 explan

0.63

 disadvant

0.56

 Beir

0.55

 coincidence

0.54

ECA

0.53

 indictment

0.52

 shenan

0.52

brainer

0.50

gie

0.50

Activations Density 1.078%

adjectives and phrases expressing opinions or evaluations

statements expressing serious concerns or issues

No Comments

No Known Activations