INDEX

Explanations

locations or positions of objects or entities

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate positioning or placement

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

ebook

-0.90

externalActionCode

-0.90

vous

-0.89

core

-0.82

ãĥ«

-0.77

displayText

-0.75

bie

-0.74

cause

-0.74

bling

-0.73

oa

-0.73

POSITIVE LOGITS

 positioned

0.99

 positioning

0.81

senal

0.81

 misunder

0.75

nesday

0.74

eering

0.74

 position

0.74

 tremend

0.73

atoon

0.72

entin

0.72

Activations Density 0.029%

locations or positions of objects or entities

phrases that indicate positioning or placement

No Comments

No Known Activations