INDEX

Explanations

phrases indicating statistical significance and variables in research contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.8.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

ENG

-0.71

worldly

-0.71

risis

-0.69

tsy

-0.69

stanbul

-0.66

inav

-0.66

burgh

-0.66

 earthqu

-0.63

Reading

-0.63

leck

-0.62

POSITIVE LOGITS

 excludes

0.84

 exclude

0.84

 unable

0.81

 exclusion

0.78

 cannot

0.77

 unus

0.72

 excluded

0.72

 preferring

0.70

 nevertheless

0.70

 avoids

0.68

Activations Density 0.858%

phrases indicating statistical significance and variables in research contexts

No Comments

No Known Activations