INDEX

Explanations

phrases related to physical locations or territories

oai_token-act-pair · gpt-3.5-turbo

references to territorial claims or conflicts

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

istries

-0.85

ems

-0.81

esson

-0.80

izations

-0.79

ustainable

-0.77

essee

-0.77

legraph

-0.76

omnia

-0.76

ynchronous

-0.75

yss

-0.74

POSITIVE LOGITS

 turf

0.99

cised

0.81

mpire

0.76

header

0.75

dab

0.73

cers

0.69

 Doodle

0.68

lette

0.67

toe

0.66

 Blades

0.66

Activations Density 0.042%

phrases related to physical locations or territories

references to territorial claims or conflicts

No Comments

No Known Activations