INDEX

Explanations

the phrase "turns out to be"

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "turns out to be."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anwhile

-0.75

ongyang

-0.74

ÃĥÃĤ

-0.70

iasco

-0.67

ipeg

-0.66

rones

-0.62

ablishment

-0.62

afia

-0.61

 Borders

-0.60

 Cran

-0.60

POSITIVE LOGITS

leeve

0.95

 able

0.90

hemoth

0.82

 regarded

0.79

heading

0.74

 influenced

0.72

Done

0.72

rael

0.72

 considered

0.71

 helpful

0.71

Activations Density 0.123%

the phrase "turns out to be"

instances of the phrase "turns out to be."

No Comments

No Known Activations