INDEX

Explanations

news-related terms, possibly related to financial news

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "WS" or similar patterns in the text

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

Ô

-0.79

ments

-0.78

mented

-0.76

orescent

-0.75

 Belg

-0.69

 ILCS

-0.68

ment

-0.67

leased

-0.65

ly

-0.64

staking

-0.62

POSITIVE LOGITS

WS

1.18

WS

1.13

pace

0.86

GF

0.84

ullivan

0.83

kins

0.82

ettings

0.80

nesday

0.80

ignt

0.78

omnia

0.78

Activations Density 0.006%

news-related terms, possibly related to financial news

occurrences of the term "WS" or similar patterns in the text

No Comments

No Known Activations