INDEX

Explanations

instances of the word "that" followed by specific information or actions

oai_token-act-pair · gpt-3.5-turbo

repeated references to specific entities or statements within an argument

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

 Roose

-0.78

OPS

-0.73

rollers

-0.69

 Loch

-0.66

 Payton

-0.66

 Samson

-0.66

lv

-0.66

 Bots

-0.65

OSH

-0.64

jack

-0.64

POSITIVE LOGITS

 particular

1.02

 same

0.93

 person

0.84

 affected

0.77

 entails

0.74

 organism

0.74

 entity

0.74

 subset

0.74

 violates

0.73

 individual

0.72

Activations Density 0.216%

instances of the word "that" followed by specific information or actions

repeated references to specific entities or statements within an argument

No Comments

No Known Activations