INDEX

Explanations

technical instructions or codes within text

oai_token-act-pair · gpt-3.5-turbo

technical terms and identifiers related to digital access and security mechanisms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

abama

-0.89

 coli

-0.75

SEA

-0.66

 Mechdragon

-0.66

 hosp

-0.66

 Pilgrim

-0.65

 Veil

-0.59

 Nationwide

-0.59

 Trooper

-0.58

 COUR

-0.58

POSITIVE LOGITS

 bearer

0.74

chain

0.74

Format

0.72

bery

0.69

finger

0.69

olkien

0.67

XL

0.67

itiveness

0.67

 crochet

0.67

acters

0.66

Activations Density 0.601%

technical instructions or codes within text

technical terms and identifiers related to digital access and security mechanisms

No Comments

No Known Activations