INDEX

Explanations

phrases indicating taking action or making progress towards a goal

oai_token-act-pair · gpt-3.5-turbo

phrases related to transitions or changes in position or status

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

 claimed

-0.77

 exploded

-0.72

 objected

-0.71

iked

-0.69

 emailed

-0.69

anded

-0.69

ochem

-0.68

 persisted

-0.66

 contributed

-0.66

 raged

-0.65

POSITIVE LOGITS

 stepping

3.07

 jumping

2.12

 launching

1.62

 quitting

1.46

 leaping

1.39

 hopping

1.38

 dumping

1.36

 skipping

1.34

 switching

1.34

 joining

1.33

Activations Density 0.028%

phrases indicating taking action or making progress towards a goal

phrases related to transitions or changes in position or status

No Comments

No Known Activations