INDEX

Explanations

specific lists, recommendations, or instructions within text

oai_token-act-pair · gpt-3.5-turbo

lists of favorites or rankings, particularly in relation to films or similar collections

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

quickShipAvailable

-0.25

=-=-

-0.23

quote

-0.22

omorphic

-0.22

ocre

-0.22

erity

-0.21

bsp

-0.21

 sidx

-0.21

inges

-0.21

daq

-0.21

POSITIVE LOGITS

Haz

0.17

 Moss

0.17

 Desert

0.17

 embr

0.17

 ethn

0.16

Gur

0.16

 Zoro

0.16

 Garrison

0.16

Hel

0.16

 loads

0.16

Activations Density 9.096%

specific lists, recommendations, or instructions within text

lists of favorites or rankings, particularly in relation to films or similar collections

No Comments

No Known Activations