INDEX

Explanations

phrases emphasizing correctness or errors

oai_token-act-pair · gpt-3.5-turbo

words that indicate a sense of correctness or appropriateness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ļéĨĴ

-0.82

ntil

-0.72

Ĭ

-0.71

ĳ

-0.69

ADRA

-0.69

antam

-0.69

ibal

-0.61

quished

-0.60

³

-0.60

Interstitial

-0.60

POSITIVE LOGITS

 messenger

0.74

 money

0.73

 wallets

0.73

 livelihood

0.72

 mindset

0.70

 destination

0.69

 mentality

0.68

 guiActiveUnfocused

0.68

 execution

0.67

 Fortune

0.67

Activations Density 0.520%

phrases emphasizing correctness or errors

words that indicate a sense of correctness or appropriateness

No Comments

No Known Activations