INDEX

Explanations

terms that suggest significant events or concepts within structured contexts, particularly in legal, social, or competitive scenarios

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

govtrack

-0.86

VIDIA

-0.85

omething

-0.83

 disadvant

-0.81

iHUD

-0.80

everal

-0.79

UCT

-0.78

Ire

-0.75

ECA

-0.75

CLASSIFIED

-0.72

POSITIVE LOGITS

ttes

0.77

ld

0.71

aux

0.68

quin

0.68

rence

0.66

redo

0.64

});

0.63

sam

0.63

ller

0.62

pt

0.62

Activations Density 0.078%

terms that suggest significant events or concepts within structured contexts, particularly in legal, social, or competitive scenarios

No Comments

No Known Activations