INDEX

Explanations

expressions indicating difficulty, doubt, or speculation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating difficulty or challenge

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

 embodiments

-0.62

 Flags

-0.61

 Reborn

-0.58

 bount

-0.58

 srfAttach

-0.58

 TAMADRA

-0.57

 Settlement

-0.57

 Stories

-0.57

Lif

-0.57

 Passage

-0.57

POSITIVE LOGITS

 imagine

1.41

 quantify

1.27

 reconcile

1.26

 envision

1.19

 justify

1.18

 conceive

1.18

 believe

1.13

 argue

1.12

 comprehend

1.10

 discern

1.07

Activations Density 0.072%

expressions indicating difficulty, doubt, or speculation

phrases indicating difficulty or challenge

No Comments

No Known Activations