INDEX

Explanations

references related to politics, specifically mentioning the Kremlin

oai_token-act-pair · gpt-3.5-turbo

references to the Kremlin and its influence

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

ORGE

-0.84

Thom

-0.83

 Kear

-0.79

âĢ¢âĢ¢âĢ¢âĢ¢

-0.73

matter

-0.72

 Kindle

-0.72

 Hath

-0.71

bee

-0.70

sbm

-0.69

Haw

-0.69

POSITIVE LOGITS

 Kremlin

0.95

 dictatorship

0.92

 rall

0.90

 dictator

0.88

achev

0.80

aganda

0.78

 Jinping

0.77

iets

0.77

 annexed

0.74

 decree

0.74

Activations Density 0.006%

references related to politics, specifically mentioning the Kremlin

references to the Kremlin and its influence

No Comments

No Known Activations