INDEX

Explanations

words related to things that are considered unimportant or insignificant in a given context

oai_token-act-pair · gpt-3.5-turbo

references to triviality and the nature of inconsequential matters

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

angelo

-0.74

elson

-0.71

elsen

-0.69

asio

-0.68

 counselor

-0.67

pa

-0.66

ario

-0.66

 counselors

-0.65

hound

-0.64

andr

-0.62

POSITIVE LOGITS

 innocuous

0.99

enance

0.90

ufact

0.89

lihood

0.85

theless

0.83

 mortals

0.82

istically

0.81

 inconven

0.80

 trivial

0.79

phrine

0.79

Activations Density 0.040%

words related to things that are considered unimportant or insignificant in a given context

references to triviality and the nature of inconsequential matters

No Comments

No Known Activations