INDEX

Explanations

non-English characters or phrases

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols indicating a translation or notation context

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

apse

-0.73

enburg

-0.73

stadt

-0.73

ileaks

-0.73

ignment

-0.68

ako

-0.67

umbledore

-0.67

ouched

-0.66

aries

-0.66

uminati

-0.65

POSITIVE LOGITS

âĸĵ

1.00

DIT

0.95

iod

0.83

¡

0.82

BLE

0.80

ĵ

0.78

uses

0.76

urs

0.73

CEPT

0.73

USE

0.72

Activations Density 0.008%

non-English characters or phrases

special characters or symbols indicating a translation or notation context

No Comments

No Known Activations