INDEX

Explanations

references to interactions or behaviors involving helping, influence, or manipulation

oai_token-act-pair · gpt-3.5-turbo

references to power dynamics and relationships in social contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Hopefully

-0.73

chester

-0.69

Ultra

-0.66

fortunately

-0.65

almost

-0.65

 EVEN

-0.64

Ĥª

-0.64

redibly

-0.62

Pretty

-0.62

âĤ¬

-0.61

POSITIVE LOGITS

 oneself

1.11

 spouse

0.82

 omission

0.79

 periphery

0.78

 duty

0.77

 infancy

0.75

effic

0.75

 superiors

0.74

 perman

0.74

 nonex

0.73

Activations Density 0.310%

references to interactions or behaviors involving helping, influence, or manipulation

references to power dynamics and relationships in social contexts

No Comments

No Known Activations