INDEX

Explanations

terms related to a specific person's name followed by a distinctive number, likely indicative of something significant related to that person

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "den."

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

 scenery

-0.68

 synergy

-0.64

 Fasc

-0.63

++;

-0.62

 bats

-0.61

 succession

-0.61

 beams

-0.59

 fasc

-0.59

 golden

-0.58

 railways

-0.58

POSITIVE LOGITS

den

4.62

Den

1.90

DEN

1.86

don

1.66

Den

1.64

den

1.55

DEN

1.36

der

1.35

die

1.35

dan

1.33

Activations Density 0.004%

terms related to a specific person's name followed by a distinctive number, likely indicative of something significant related to that person

occurrences of the word "den."

No Comments

No Known Activations