INDEX

Explanations

proper nouns like names of people and places

oai_token-act-pair · gpt-3.5-turbo

alphanumeric sequences and symbols, potentially indicating technical or coding information

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

 Turing

-0.59

ruary

-0.57

shire

-0.55

 Skinner

-0.50

 ACTIONS

-0.50

 unse

-0.48

 forgotten

-0.48

 ceremon

-0.47

 confidentiality

-0.46

tein

-0.46

POSITIVE LOGITS

ãĥĺãĥ©

0.72

agi

0.67

soDeliveryDate

0.66

drm

0.65

 Gujar

0.53

ãĥŁ

0.53

Ï

0.52

achu

0.52

allery

0.51

itars

0.51

Activations Density 1.401%

proper nouns like names of people and places

alphanumeric sequences and symbols, potentially indicating technical or coding information

No Comments

No Known Activations