INDEX

Explanations

words that appear to be in a different language

oai_token-act-pair · gpt-3.5-turbo

foreign or non-English words and characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Companies

-0.69

Sharp

-0.67

-0.65

Neal

-0.63

 Jarrett

-0.62

Bill

-0.62

 Banks

-0.62

 claimant

-0.60

 Objective

-0.60

POSITIVE LOGITS

usalem

1.03

liga

0.97

Ã§a

0.94

ument

0.93

olog

0.93

asta

0.92

quer

0.90

terior

0.85

umat

0.83

celona

0.82

Activations Density 0.168%

words that appear to be in a different language

foreign or non-English words and characters

No Comments

No Known Activations