INDEX

Explanations

text related to redacted or hidden information in documents

oai_token-act-pair · gpt-3.5-turbo

terms related to legal proceedings and government record management

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

 masc

-0.78

cil

-0.74

 charismatic

-0.71

ominated

-0.70

Econom

-0.68

 Swanson

-0.68

ombat

-0.68

atu

-0.67

atos

-0.67

acan

-0.65

POSITIVE LOGITS

 FOIA

2.05

 redacted

1.67

 subpoen

1.52

 declass

1.49

 subpoena

1.44

 confidentiality

1.44

 documents

1.42

 archive

1.37

 records

1.34

 archives

1.33

Activations Density 0.419%

text related to redacted or hidden information in documents

terms related to legal proceedings and government record management

No Comments

No Known Activations