INDEX

Explanations

mentions of candles and candle-related activities

oai_token-act-pair · gpt-3.5-turbo

references to candles and related themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

doms

-0.79

ordan

-0.73

Reviewer

-0.68

iago

-0.66

 havoc

-0.65

 Subjects

-0.65

laus

-0.64

ese

-0.64

omal

-0.64

omo

-0.63

POSITIVE LOGITS

 candle

1.16

lit

1.08

 Candle

1.04

 candles

1.00

light

0.93

warm

0.89

 lantern

0.89

 vigil

0.85

holder

0.85

pipe

0.85

Activations Density 0.025%

mentions of candles and candle-related activities

references to candles and related themes

No Comments

No Known Activations