INDEX

Explanations

references to TED events or talks

oai_token-act-pair · gpt-3.5-turbo

references to TEDx events and talks

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

istics

-0.85

alogy

-0.77

antine

-0.71

 Shinra

-0.69

dinand

-0.69

lein

-0.68

istration

-0.67

alog

-0.67

heses

-0.62

ardless

-0.62

POSITIVE LOGITS

MED

0.88

ulously

0.80

TED

0.78

quickShipAvailable

0.75

xs

0.74

Interstitial

0.73

endum

0.72

è¯

0.70

odied

0.69

PROV

0.68

Activations Density 0.027%

references to TED events or talks

references to TEDx events and talks

No Comments

No Known Activations