INDEX

Explanations

superlatives or things related to the highest level of something

oai_token-act-pair · gpt-3.5-turbo

terms that indicate the highest levels or rankings in various contexts

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

icing

-0.69

redo

-0.66

opard

-0.64

icer

-0.63

agra

-0.63

vous

-0.62

kay

-0.62

izon

-0.61

ching

-0.60

rahim

-0.59

POSITIVE LOGITS

 bidder

0.91

 rated

0.88

 pinnacle

0.86

 concentrations

0.79

hest

0.79

 imaginable

0.79

 earners

0.77

 possible

0.75

 proport

0.75

 extent

0.74

Activations Density 0.018%

superlatives or things related to the highest level of something

terms that indicate the highest levels or rankings in various contexts

No Comments

No Known Activations