INDEX

Explanations

keywords related to different generations, particularly Generation 9

oai_token-act-pair · gpt-3.5-turbo

references to generational labels or cohorts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rig

-0.81

 haunted

-0.77

 hanged

-0.74

 hosp

-0.73

 nerve

-0.70

rob

-0.70

ounty

-0.69

clair

-0.69

 physically

-0.69

emb

-0.68

POSITIVE LOGITS

 Generation

1.23

 Generations

1.03

idable

0.86

 Fighters

0.81

 Cars

0.80

naire

0.77

 Equality

0.75

azi

0.75

 Genie

0.73

 Sequence

0.72

Activations Density 0.009%

keywords related to different generations, particularly Generation 9

references to generational labels or cohorts

No Comments

No Known Activations