INDEX

Explanations

repetitive mentions or actions represented by the word "again"

oai_token-act-pair · gpt-3.5-turbo

repeated mentions or events happening again

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

ãģ®éŃĶ

-0.64

:{

-0.63

Gameplay

-0.62

Contents

-0.61

iol

-0.61

IELD

-0.60

ahime

-0.59

minecraft

-0.59

scribe

-0.58

 careers

-0.58

POSITIVE LOGITS

 again

1.19

again

1.18

 Again

1.12

this

1.12

 this

1.03

Again

0.90

 albeit

0.89

 another

0.83

another

0.82

same

0.81

Activations Density 0.362%

repetitive mentions or actions represented by the word "again"

repeated mentions or events happening again

No Comments

No Known Activations