INDEX

Explanations

personal or confidential information being revealed or shared

oai_token-act-pair · gpt-3.5-turbo

past-tense verbs or participles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

heter

-0.63

CHA

-0.62

OVA

-0.60

RP

-0.59

çĶ°

-0.58

 Carbuncle

-0.56

 suspects

-0.56

-0.55

NG

-0.55

Hu

-0.53

POSITIVE LOGITS

ided

1.09

escription

0.92

 Dhabi

0.89

iding

0.87

ide

0.87

oggle

0.83

own

0.83

ãĥ¼ãĤ¯

0.82

IDE

0.78

SourceFile

0.76

Activations Density 0.010%

personal or confidential information being revealed or shared

past-tense verbs or participles

No Comments

No Known Activations