INDEX

Explanations

mentions of absence or missing presence

oai_token-act-pair · gpt-3.5-turbo

forms of the word "absent"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Ocean

-0.87

ipel

-0.82

ombs

-0.79

Bre

-0.79

RAW

-0.79

\\\\\\\\

-0.74

ocard

-0.72

RAFT

-0.71

FK

-0.71

emetery

-0.70

POSITIVE LOGITS

minded

1.09

 minded

0.97

alus

0.71

soever

0.70

boxing

0.69

 Mondays

0.68

lihood

0.68

comings

0.66

icularly

0.66

icular

0.66

Activations Density 0.011%

mentions of absence or missing presence

forms of the word "absent"

No Comments

No Known Activations