INDEX

Explanations

phrases related to courtroom proceedings and legal cases

oai_token-act-pair · gpt-3.5-turbo

instances of the word "heard" and related variations

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

iets

-0.72

oil

-0.66

 leaps

-0.65

idation

-0.64

icidal

-0.64

ifiable

-0.64

ocide

-0.63

ãĤ¦

-0.63

 mistake

-0.63

 mell

-0.62

POSITIVE LOGITS

 Hear

0.99

ropolitan

0.95

 Hearing

0.86

 hearings

0.85

 testimony

0.84

 hearing

0.84

ĸļ

0.80

NetMessage

0.75

Ear

0.72

sbm

0.72

Activations Density 0.033%

phrases related to courtroom proceedings and legal cases

instances of the word "heard" and related variations

No Comments

No Known Activations