INDEX

Explanations

commands or instructions indicating actions that can be performed

oai_token-act-pair · gpt-3.5-turbo

modal verbs indicating ability or potential actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pires

-0.76

 Hits

-0.66

 Niet

-0.65

 Closing

-0.63

bush

-0.63

Moz

-0.63

cule

-0.62

 Mour

-0.61

polit

-0.61

bred

-0.59

POSITIVE LOGITS

't

1.35

 optionally

1.19

 choose

1.17

 customize

1.06

 utilize

1.06

NOT

1.05

 skip

1.04

 omit

1.01

use

1.00

 modify

0.99

Activations Density 0.163%

commands or instructions indicating actions that can be performed

modal verbs indicating ability or potential actions

No Comments

No Known Activations