INDEX

Explanations

words related to the University of Washington

oai_token-act-pair · gpt-3.5-turbo

references to a specific location, indicated by the term "Wa."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

rious

-0.74

sis

-0.74

lycer

-0.72

displayText

-0.71

xual

-0.70

erous

-0.69

sticks

-0.69

tics

-0.67

CPC

-0.65

rations

-0.65

POSITIVE LOGITS

velength

1.45

Wa

1.04

aii

0.94

apon

0.91

ivers

0.90

Wa

0.89

veland

0.89

heed

0.88

pless

0.88

ipeg

0.87

Activations Density 0.005%

words related to the University of Washington

references to a specific location, indicated by the term "Wa."

No Comments

No Known Activations