INDEX

Explanations

words related to investigative actions and legal proceedings

oai_token-act-pair · gpt-3.5-turbo

frequent auxiliary or linking verbs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ilaterally

-0.64

edIn

-0.61

æµ

-0.56

ãĥ¼ãĥĨ

-0.55

 Moroc

-0.54

emale

-0.53

 thous

-0.53

chuk

-0.52

 Seym

-0.51

uese

-0.51

POSITIVE LOGITS

 WATCHED

0.53

 largeDownload

0.51

CRIPTION

0.51

ANS

0.49

zbollah

0.48

SELECT

0.48

 âĢº

0.47

 SERV

0.47

Interested

0.47

EVE

0.46

Activations Density 1.606%

words related to investigative actions and legal proceedings

frequent auxiliary or linking verbs

No Comments

No Known Activations