INDEX

Explanations

references to being in jail or facing imprisonment

oai_token-act-pair · gpt-3.5-turbo

terminology related to incarceration and imprisonment

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

 plurality

-0.69

ulner

-0.65

quart

-0.64

onal

-0.64

ween

-0.63

ophe

-0.62

rex

-0.62

cific

-0.61

ombs

-0.61

ript

-0.60

POSITIVE LOGITS

 limbo

0.72

 Reloaded

0.70

 ..........

0.69

pedia

0.69

 hearings

0.68

Ø©

0.66

 awaiting

0.64

 proceedings

0.61

 Hearing

0.59

 dehydration

0.59

Activations Density 0.077%

references to being in jail or facing imprisonment

terminology related to incarceration and imprisonment

No Comments

No Known Activations