INDEX

Explanations

references to the word "Mercury" in a text

oai_token-act-pair · gpt-3.5-turbo

mentions of "Mercury" and related terms

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

-0.86

Redditor

-0.77

animate

-0.74

itcher

-0.74

dat

-0.68

eps

-0.67

tk

-0.66

eks

-0.66

ãĥł

-0.65

olkien

-0.65

POSITIVE LOGITS

 Mercury

1.19

iod

0.88

 mosqu

0.80

 Seal

0.75

 certs

0.73

 tremend

0.73

£ı

0.72

oun

0.69

 seals

0.69

urnal

0.68

Activations Density 0.006%

references to the word "Mercury" in a text

mentions of "Mercury" and related terms

No Comments

No Known Activations