INDEX

Explanations

adjectives that describe the intensity of something

oai_token-act-pair · gpt-3.5-turbo

emphatic expressions of opinion or intensity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

antry

-0.86

 Altern

-0.72

icipated

-0.69

 citation

-0.63

heid

-0.63

arthed

-0.61

 Receiver

-0.61

engers

-0.61

Previously

-0.61

amide

-0.59

POSITIVE LOGITS

 nice

0.91

 neat

0.91

 cool

0.91

 messed

0.89

 pissed

0.87

nice

0.87

 nasty

0.86

 freaking

0.84

 interesting

0.83

 scary

0.82

Activations Density 0.065%

adjectives that describe the intensity of something

emphatic expressions of opinion or intensity

No Comments

No Known Activations