INDEX

Explanations

proper nouns related to individuals or characters

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

Negative Logits

 Rollins

-0.71

nikov

-0.65

SHIP

-0.64

sworth

-0.62

 Bauer

-0.61

daq

-0.61

kefeller

-0.60

nostic

-0.59

Neh

-0.59

DAQ

-0.59

POSITIVE LOGITS

union

0.83

omp

0.78

vent

0.76

Ãª

0.72

inter

0.68

actionDate

0.67

ompl

0.67

omm

0.66

river

0.66

arbon

0.65

Activations Density 0.267%

proper nouns related to individuals or characters

No Known Activations