INDEX

Explanations

references to underground locations and activities

oai_token-act-pair · gpt-3.5-turbo

references to actions involving movement and discovery

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

à¸

-0.85

 Vanessa

-0.83

actionDate

-0.79

 Riot

-0.77

 Synd

-0.75

 Reck

-0.73

rencies

-0.72

 Proud

-0.71

 Clarkson

-0.71

 Alic

-0.71

POSITIVE LOGITS

 tunnels

1.97

 tunnel

1.90

 cavern

1.81

 hole

1.73

 caves

1.70

 cave

1.65

 subter

1.61

 holes

1.54

 excavation

1.54

 underground

1.50

Activations Density 0.342%

references to underground locations and activities

references to actions involving movement and discovery

No Comments

No Known Activations