INDEX

Explanations

texts related to arguments and debates

oai_token-act-pair · gpt-3.5-turbo

phrases related to presenting and discussing arguments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 PHOTO

-0.70

 Carbuncle

-0.67

 Nare

-0.62

 Atomic

-0.60

idays

-0.60

 Enterprises

-0.59

 Eclipse

-0.58

 curfew

-0.58

ookie

-0.58

hit

-0.57

POSITIVE LOGITS

ative

1.26

 against

1.06

uments

1.06

abl

1.05

ument

0.96

 persu

0.96

Against

0.93

against

0.90

ation

0.88

ator

0.87

Activations Density 0.046%

texts related to arguments and debates

phrases related to presenting and discussing arguments

No Comments

No Known Activations