INDEX

Explanations

sentences requesting comments or responses

oai_token-act-pair · gpt-3.5-turbo

instances of comments or statements made in response to requests for commentary

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

 tremend

-0.86

 rejuven

-0.80

 mosqu

-0.79

isable

-0.77

 painfully

-0.77

 dracon

-0.77

 regener

-0.76

 ascend

-0.74

 unstoppable

-0.73

 millenn

-0.73

POSITIVE LOGITS

 Neither

1.27

<|endoftext|>

1.26

 However

1.18

 Officials

1.15

 Nonetheless

1.12

 Earlier

1.12

 Â®

1.11

 Regardless

1.11

 Asked

1.11

 Nevertheless

1.07

Activations Density 0.241%

sentences requesting comments or responses

instances of comments or statements made in response to requests for commentary

No Comments

No Known Activations