INDEX

Explanations

phrases expressing clarity or candor

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing clarity and straightforwardness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ources

-0.70

issance

-0.64

 Combine

-0.64

urn

-0.62

naires

-0.59

ighting

-0.58

ERE

-0.58

orer

-0.57

 pockets

-0.57

 Erie

-0.56

POSITIVE LOGITS

 blunt

1.15

 frank

1.12

 honest

1.04

 unequiv

0.96

 clear

0.96

 explicit

0.90

 simple

0.87

 precise

0.86

 truthful

0.85

simple

0.85

Activations Density 0.092%

phrases expressing clarity or candor

phrases emphasizing clarity and straightforwardness

No Comments

No Known Activations