INDEX

Explanations

words related to talent or talented individuals

oai_token-act-pair · gpt-3.5-turbo

references to talent and skill

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ocide

-0.68

ety

-0.65

iso

-0.65

ep

-0.64

resp

-0.61

AMA

-0.61

 Conquest

-0.60

 Statement

-0.60

ombat

-0.59

bas

-0.58

POSITIVE LOGITS

 talented

3.35

 skilled

1.93

 talent

1.81

 gifted

1.66

 inexperienced

1.56

 talents

1.50

 knowledgeable

1.37

 versatile

1.34

 competent

1.32

 charismatic

1.32

Activations Density 0.009%

words related to talent or talented individuals

references to talent and skill

No Comments

No Known Activations