INDEX

Explanations

phrases that signify legislative actions or processes

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

andowski

-0.77

 caut

-0.75

 disabling

-0.72

netflix

-0.71

 disabled

-0.67

ffff

-0.66

 disqual

-0.65

BIP

-0.64

PIN

-0.63

govtrack

-0.63

POSITIVE LOGITS

 accordance

0.70

 conjunction

0.69

erie

0.66

 Soldiers

0.64

aneous

0.64

 relation

0.63

hoff

0.62

 batches

0.61

 lieu

0.59

 response

0.59

Activations Density 0.136%

phrases that signify legislative actions or processes

No Comments

No Known Activations