INDEX

Explanations

the word "gratifying" and related variations

oai_token-act-pair · gpt-3.5-turbo

expressions of gratitude or positive recognition

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Formation

-0.69

 Regulation

-0.64

 FIGHT

-0.63

 Alic

-0.63

 Prediction

-0.63

 Universities

-0.62

 Patriarch

-0.62

 Construct

-0.62

 shortened

-0.61

 Resistance

-0.60

POSITIVE LOGITS

uitous

1.55

uit

1.19

uity

1.09

 grat

1.08

iously

1.01

ified

0.91

animous

0.89

ieved

0.89

uously

0.88

iless

0.88

Activations Density 0.016%

the word "gratifying" and related variations

expressions of gratitude or positive recognition

No Comments

No Known Activations