INDEX

Explanations

phrases related to speaking or sharing information under specific conditions

oai_token-act-pair · gpt-3.5-turbo

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

Ļ½

-0.80

 discont

-0.61

ģĸ

-0.60

ories

-0.59

ptoms

-0.58

 spir

-0.54

 impacted

-0.54

pex

-0.54

 impacting

-0.53

Cue

-0.52

POSITIVE LOGITS

 anonymity

1.59

 confidentiality

1.00

 privacy

0.88

 secrecy

0.86

 anonym

0.76

ullah

0.75

 permissions

0.74

 exemptions

0.72

 permission

0.70

 encryption

0.70

Activations Density 0.035%

phrases related to speaking or sharing information under specific conditions

No Comments

No Known Activations