INDEX

Explanations

phrases related to putting something on record

oai_token-act-pair · gpt-3.5-turbo

references to being "on the record" in discussions or statements

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

 Flavoring

-0.71

incinn

-0.71

aturation

-0.69

issan

-0.65

ocate

-0.65

iant

-0.64

 Franch

-0.64

cius

-0.64

vec

-0.63

creen

-0.62

POSITIVE LOGITS

 record

0.78

 linkage

0.76

record

0.75

 labels

0.71

 Telegram

0.67

Hay

0.65

 records

0.63

catentry

0.62

 refere

0.62

raid

0.62

Activations Density 0.014%

phrases related to putting something on record

references to being "on the record" in discussions or statements

No Comments

No Known Activations