INDEX

Explanations

references to specific historical years

oai_token-act-pair · gpt-3.5-turbo

references to specific historical years

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

•Embedding Space Digit Features

No Comments

Negative Logits

*/(

-0.77

igi

-0.68

ony

-0.67

 Corker

-0.66

FILE

-0.65

FH

-0.62

McN

-0.61

 Donovan

-0.61

arr

-0.60

çĶŁ

-0.60

POSITIVE LOGITS

2.84

2.64

2.56

2.54

2.49

2.41

2.34

2.33

2.28

Activations Density 0.028%

references to specific historical years

references to specific historical years

No Comments

No Known Activations