INDEX

Explanations

phrases related to mistakes, errors, flaws, and vulnerabilities

oai_token-act-pair · gpt-3.5-turbo

terms associated with flaws and failures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

avers

-0.74

ignt

-0.73

 peanuts

-0.70

umption

-0.67

esta

-0.67

addle

-0.67

omers

-0.66

pta

-0.66

angel

-0.64

 Primordial

-0.64

POSITIVE LOGITS

lessly

1.55

ously

1.09

lessness

1.08

ably

1.04

nah

0.95

iously

0.83

 flaw

0.82

itsu

0.82

tie

0.80

lier

0.79

Activations Density 0.014%

phrases related to mistakes, errors, flaws, and vulnerabilities

terms associated with flaws and failures

No Comments

No Known Activations