INDEX

Explanations

verbs relating to knowledge or understanding

oai_token-act-pair · gpt-3.5-turbo

the word "know" in various contexts, suggesting a focus on awareness or knowledge

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

phrine

-0.90

 earthqu

-0.76

isco

-0.76

oples

-0.74

ItemTracker

-0.74

conservancy

-0.71

packages

-0.70

interstitial

-0.69

uld

-0.69

anmar

-0.67

POSITIVE LOGITS

ledged

1.05

ledge

0.93

lege

0.81

 KNOW

0.81

how

0.76

HOW

0.76

hent

0.74

how

0.72

 Nare

0.70

ariat

0.70

Activations Density 0.060%

verbs relating to knowledge or understanding

the word "know" in various contexts, suggesting a focus on awareness or knowledge

No Comments

No Known Activations