INDEX

Explanations

numeric values or identifiers, possibly related to technology or specific products

oai_token-act-pair · gpt-3.5-turbo

specific numerical values and the name "Sharif."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nces

-0.80

NetMessage

-0.80

illac

-0.77

asures

-0.73

gas

-0.72

mble

-0.70

ktop

-0.69

lyak

-0.67

lli

-0.67

onym

-0.66

POSITIVE LOGITS

 Carbuncle

0.93

0.87

smith

0.79

 Sharif

0.76

ship

0.75

abis

0.74

itudinal

0.73

ual

0.70

rontal

0.70

icial

0.70

Activations Density 0.042%

numeric values or identifiers, possibly related to technology or specific products

specific numerical values and the name "Sharif."

No Comments

No Known Activations