INDEX

Explanations

specific phrases or keywords related to instructions or prompts

oai_token-act-pair · gpt-3.5-turbo

sentence starters or common phrases indicative of dialogue and narrative structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

}}

-0.49

[|

-0.47

 referen

-0.46

idated

-0.46

ebook

-0.45

arsity

-0.44

 CLSID

-0.44

 Moroc

-0.44

ä»£

-0.44

 Polo

-0.44

POSITIVE LOGITS

etheless

0.69

ktop

0.64

mosp

0.64

resa

0.62

swers

0.62

xiety

0.59

zbollah

0.58

Voice

0.56

jriwal

0.54

notations

0.54

Activations Density 0.589%

specific phrases or keywords related to instructions or prompts

sentence starters or common phrases indicative of dialogue and narrative structure

No Comments

No Known Activations