INDEX

Explanations

phrases related to cans

oai_token-act-pair · gpt-3.5-turbo

references to containers, particularly cans

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

RESULTS

-0.82

Tut

-0.74

 Knight

-0.68

doms

-0.67

 Fighters

-0.63

Known

-0.61

 handshake

-0.61

Gameplay

-0.60

ORGE

-0.59

 Ascend

-0.59

POSITIVE LOGITS

 cans

1.17

atell

1.11

vas

0.96

alogue

0.96

alys

0.92

ohydrate

0.92

keyes

0.92

't

0.92

keye

0.85

thur

0.82

Activations Density 0.013%

phrases related to cans

references to containers, particularly cans

No Comments

No Known Activations