INDEX

Explanations

concerns or worries expressed in text

oai_token-act-pair · gpt-3.5-turbo

recurring themes of concern

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

RIP

-0.77

OUP

-0.76

Sort

-0.74

redits

-0.73

haw

-0.72

 salute

-0.71

 toast

-0.68

 Doodle

-0.68

å§

-0.68

 precincts

-0.64

POSITIVE LOGITS

severe

0.80

cens

0.78

 developing

0.75

 preventing

0.73

 mounting

0.72

 inadequ

0.72

 excessive

0.71

 undue

0.71

 litigation

0.71

 heightened

0.70

Activations Density 0.134%

concerns or worries expressed in text

recurring themes of concern

No Comments

No Known Activations