INDEX

Explanations

references to specific publications, particularly "The Economist."

oai_token-act-pair · gpt-3.5-turbo

mentions of the publication "The Economist" and related scientific terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

romy

-0.79

illon

-0.73

une

-0.71

oÄŁ

-0.71

etting

-0.70

ammy

-0.70

ificial

-0.69

oine

-0.68

illas

-0.67

ensions

-0.67

POSITIVE LOGITS

 Economist

0.97

 Intelligence

0.97

 Reader

0.82

 Perspect

0.79

ãĥ³ãĤ¸

0.78

vier

0.76

 Newspaper

0.74

 Chronicle

0.74

Journal

0.72

 Magazine

0.71

Activations Density 0.020%

references to specific publications, particularly "The Economist."

mentions of the publication "The Economist" and related scientific terms

No Comments

No Known Activations