INDEX

Explanations

mentions of the word "Huron"

oai_token-act-pair · gpt-3.5-turbo

references to the word "Hur."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ancial

-0.77

nesota

-0.75

 Skydragon

-0.70

atown

-0.69

ħĭ

-0.67

ä»£

-0.67

istration

-0.65

Appearance

-0.65

inition

-0.64

 intensive

-0.64

POSITIVE LOGITS

Hur

0.92

ricanes

0.88

ving

0.87

hur

0.87

ves

0.86

witz

0.85

ley

0.82

Hur

0.82

acan

0.81

rah

0.79

Activations Density 0.008%

mentions of the word "Huron"

references to the word "Hur."

No Comments

No Known Activations