INDEX

Explanations

sequences with the specific string "GR" in them

oai_token-act-pair · gpt-3.5-turbo

occurrences of the acronym "GR" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

eers

-0.81

gerald

-0.74

ters

-0.74

manship

-0.73

ovic

-0.70

que

-0.70

quo

-0.69

 Wand

-0.68

epad

-0.67

stra

-0.66

POSITIVE LOGITS

APH

1.25

izont

1.06

ADE

1.03

ANT

0.99

ASS

0.97

ains

0.93

ACE

0.93

ENE

0.92

OUND

0.92

ATOR

0.89

Activations Density 0.004%

sequences with the specific string "GR" in them

occurrences of the acronym "GR" and related terms

No Comments

No Known Activations