INDEX

Explanations

references to careers and seasons in sports

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

xual

-0.73

andom

-0.69

agents

-0.67

 Mehran

-0.62

esp

-0.61

 Greenberg

-0.61

chnology

-0.61

gered

-0.60

ihad

-0.59

isks

-0.58

POSITIVE LOGITS

high

1.28

 highs

1.22

 lows

1.07

 high

1.04

low

1.03

low

0.93

best

0.91

 HIGH

0.86

long

0.83

leading

0.81

Activations Density 0.645%

references to careers and seasons in sports

No Comments

No Known Activations