INDEX

Explanations

words related to specific technical concepts or jargon, possibly related to computer science and analytics

oai_token-act-pair · gpt-3.5-turbo

references to accents and their effects 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Coastal

-0.68

tails

-0.62

phasis

-0.60

hani

-0.60

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.60

rador

-0.59

illac

-0.59

yrs

-0.59

cca

-0.58

ppa

-0.58

POSITIVE LOGITS

 etiquette

0.83

 cube

0.65

cub

0.61

 experien

0.61

 ceremon

0.61

 cubes

0.61

 pals

0.61

 nodd

0.60

orb

0.59

ipedia

0.59

Activations Density 1.097%

words related to specific technical concepts or jargon, possibly related to computer science and analytics

references to accents and their effects in various contexts

No Comments

No Known Activations