INDEX

Explanations

terms related to legal penalties and offenses

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

otin

-0.75

leck

-0.71

 inher

-0.65

 visitors

-0.65

hend

-0.64

 holders

-0.60

roman

-0.59

 claimants

-0.59

 wield

-0.59

fman

-0.59

POSITIVE LOGITS

GGGG

0.91

externalActionCode

0.67

 taco

0.62

asted

0.61

'';

0.60

UCH

0.60

ī

0.59

 Array

0.58

isites

0.58

Cor

0.58

Activations Density 0.132%

terms related to legal penalties and offenses

No Comments

No Known Activations