INDEX

Explanations

verbs that signify the possible impact of actions along with terms associated with research and development

oai_token-act-pair · gpt-3.5-turbo

phrases indicating potential outcomes or future actions

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

 hates

-0.77

 Died

-0.77

 muttered

-0.74

raped

-0.74

 complains

-0.73

 cursing

-0.72

 whining

-0.72

said

-0.72

bugs

-0.71

running

-0.71

POSITIVE LOGITS

 help

1.44

 enhance

1.42

 enable

1.41

 facilitate

1.38

 encourage

1.34

 assist

1.34

 empower

1.33

 pave

1.32

 motivate

1.31

aid

1.30

Activations Density 0.219%

verbs that signify the possible impact of actions along with terms associated with research and development

phrases indicating potential outcomes or future actions

No Comments

No Known Activations