INDEX

Explanations

references to contributions and authorship in academic contexts

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

Negative Logits

cffff

-0.86

ierrez

-0.73

________________

-0.71

gerald

-0.70

ahoo

-0.67

conserv

-0.67

ãĥ¼ãĥ«

-0.67

cules

-0.66

outube

-0.65

odan

-0.64

POSITIVE LOGITS

 Overwatch

0.74

 Ethereum

0.73

NEO

0.69

 WORLD

0.69

tnc

0.62

 Committees

0.61

 Javascript

0.61

 Valentine

0.61

resses

0.61

 Machina

0.61

Activations Density 0.270%

references to contributions and authorship in academic contexts

No Known Activations