INDEX

Explanations

phrases related to successful completion or achievement of tasks

oai_token-act-pair · gpt-3.5-turbo

phrases indicating successful actions or achievements

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

 Respons

-0.73

soever

-0.73

 Actions

-0.68

 duties

-0.66

 Continued

-0.66

acted

-0.64

Next

-0.64

 Relief

-0.63

Sold

-0.63

 opin

-0.62

POSITIVE LOGITS

 convince

1.41

 achieve

1.25

 overcome

1.22

 accomplish

1.22

 persuade

1.21

 snag

1.20

 infiltrate

1.20

 evade

1.19

 survive

1.18

 muster

1.14

Activations Density 0.108%

phrases related to successful completion or achievement of tasks

phrases indicating successful actions or achievements

No Comments

No Known Activations