INDEX

Explanations

text indicating a message or link to be clicked or opened

oai_token-act-pair · gpt-3.5-turbo

the word "open"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Pound

-0.73

 Lane

-0.69

ocular

-0.65

BILITY

-0.64

\\\\\\\\\\\\\\\\

-0.64

EStreamFrame

-0.63

acular

-0.62

arettes

-0.60

BLE

-0.59

bit

-0.59

POSITIVE LOGITS

 Aval

0.70

count

0.69

properties

0.67

inois

0.62

eder

0.61

Ident

0.61

 proc

0.60

reon

0.60

Bey

0.59

 Debor

0.58

Activations Density 0.000%

text indicating a message or link to be clicked or opened

the word "open"

No Comments

No Known Activations