INDEX

Explanations

words related to action and decisions

oai_token-act-pair · gpt-3.5-turbo

the word "and," indicating a focus on conjunctions and connections between ideas

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

ummer

-0.77

wic

-0.74

icol

-0.73

felt

-0.73

lain

-0.72

IFIED

-0.72

thought

-0.71

ĸļ

-0.71

standing

-0.70

urnal

-0.67

POSITIVE LOGITS

 minimize

1.39

 reduce

1.39

 eliminate

1.38

 engage

1.36

 create

1.36

 organize

1.34

 introduce

1.34

 formulate

1.33

 impose

1.33

 maximize

1.33

Activations Density 0.283%

words related to action and decisions

the word "and," indicating a focus on conjunctions and connections between ideas

No Comments

No Known Activations