INDEX

Explanations

questions starting with "Why" followed by a verb in past tense

oai_token-act-pair · gpt-3.5-turbo

questions about political action and social relevance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disg

-0.61

EStreamFrame

-0.59

 marrow

-0.58

 confir

-0.55

illet

-0.54

 redeem

-0.54

 glim

-0.53

 redeemed

-0.53

qt

-0.53

 breeze

-0.52

POSITIVE LOGITS

instead

1.10

 Matters

0.99

 instead

0.98

?:

0.94

rather

0.91

so

0.85

 anymore

0.85

 nowadays

0.80

 rather

0.80

Instead

0.80

Activations Density 0.350%

questions starting with "Why" followed by a verb in past tense

questions about political action and social relevance

No Comments

No Known Activations