INDEX

Explanations

phrases related to leadership and decision-making

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "ad" in various contexts

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

chwitz

-0.71

 vacuum

-0.64

 wreck

-0.62

 Hemisphere

-0.62

 horizont

-0.61

 Illum

-0.61

 paralyzed

-0.61

 Square

-0.61

 available

-0.60

 poles

-0.60

POSITIVE LOGITS

ad

4.09

ads

2.74

AD

2.06

adic

2.05

adh

2.00

adan

1.78

adem

1.77

adr

1.72

adi

1.70

ador

1.66

Activations Density 0.027%

phrases related to leadership and decision-making

occurrences of the word "ad" in various contexts

No Comments

No Known Activations