INDEX

Explanations

phrases indicating a call to action or urging someone to take a specific course of action

oai_token-act-pair · gpt-3.5-turbo

calls to action directed at various entities or audiences

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

quickShipAvailable

-0.94

cases

-0.65

Yep

-0.60

database

-0.60

alysed

-0.60

paren

-0.60

radio

-0.59

effects

-0.59

 spawned

-0.59

traumatic

-0.58

POSITIVE LOGITS

 reconsider

1.15

 avoid

1.07

be

1.05

 participate

1.04

 refrain

1.02

 preserve

0.99

 keep

0.99

 comply

0.98

 undertake

0.98

 emulate

0.97

Activations Density 0.148%

phrases indicating a call to action or urging someone to take a specific course of action

calls to action directed at various entities or audiences

No Comments

No Known Activations