INDEX

Explanations

organizations or entities mentioned in news articles or similar texts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating participants or entities in a list of noteworthy individuals or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

acy

-0.70

é»Ĵ

-0.69

 Mahjong

-0.67

 sorcery

-0.66

 Armageddon

-0.65

Wars

-0.63

 Reincarn

-0.62

 immortality

-0.62

 Efficiency

-0.61

cation

-0.61

POSITIVE LOGITS

 beneficiary

1.11

 culprit

1.08

 beneficiaries

1.06

 recipients

0.97

 casualties

0.92

 recipient

0.88

 casualty

0.87

 participant

0.85

 participants

0.83

 inspir

0.82

Activations Density 0.261%

organizations or entities mentioned in news articles or similar texts

phrases indicating participants or entities in a list of noteworthy individuals or groups

No Comments

No Known Activations