INDEX

Explanations

prepositions followed by verbs

oai_token-act-pair · gpt-3.5-turbo

emphatic or reinforcing words and phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

regor

-0.59

 Seym

-0.58

chuk

-0.55

Keep

-0.54

 Grind

-0.52

[*

-0.52

Ultra

-0.51

enegger

-0.51

Gear

-0.51

Swe

-0.51

POSITIVE LOGITS

ye

0.64

 Wikimedia

0.58

 hereby

0.55

 thou

0.53

 shalt

0.52

 preferably

0.52

 bearer

0.52

-)

0.51

ya

0.51

 Â£

0.50

Activations Density 0.485%

prepositions followed by verbs

emphatic or reinforcing words and phrases

No Comments

No Known Activations