INDEX

Explanations

the word "Kn" followed by a number, possibly referring to a specific entity or concept

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names or terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

quo

-0.85

dit

-0.76

ORGE

-0.73

 Uran

-0.72

AQ

-0.67

ahime

-0.64

pour

-0.64

 bubbles

-0.64

lords

-0.63

REDACTED

-0.63

POSITIVE LOGITS

itting

1.03

ocking

1.02

ivable

1.00

ows

0.99

uckle

0.98

itty

0.94

ock

0.94

uth

0.94

icker

0.92

keye

0.89

Activations Density 0.014%

the word "Kn" followed by a number, possibly referring to a specific entity or concept

mentions of specific names or terms

No Comments

No Known Activations