INDEX

Explanations

specific names or entities, in particular, ones related to journalism and technology

oai_token-act-pair · gpt-3.5-turbo

mentions of specific media organizations and key individuals associated with them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 reconc

-0.88

oke

-0.75

BuyableInstoreAndOnline

-0.71

 separ

-0.67

avorite

-0.67

 unification

-0.66

ogun

-0.64

 delim

-0.63

 simplistic

-0.61

 captcha

-0.61

POSITIVE LOGITS

 Intercept

1.11

 Greenwald

1.01

ionage

0.91

Accessory

0.89

 GOODMAN

0.85

ileaks

0.85

Guard

0.80

Leaks

0.79

 Frontier

0.77

Freedom

0.75

Activations Density 0.027%

specific names or entities, in particular, ones related to journalism and technology

mentions of specific media organizations and key individuals associated with them

No Comments

No Known Activations