INDEX

Explanations

action words relating to investigations or explorations

oai_token-act-pair · gpt-3.5-turbo

technical terms or descriptors related to integration and participation

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

 shroud

-0.72

 thereto

-0.66

 companion

-0.65

 respectively

-0.65

 âĢ¦"

-0.64

IF

-0.64

 sanctuary

-0.64

 polyg

-0.63

EVE

-0.62

".

-0.61

POSITIVE LOGITS

resa

1.09

roximately

1.06

aring

0.99

inarily

0.96

ilities

0.95

erate

0.95

autions

0.94

vertising

0.93

inct

0.92

ison

0.92

Activations Density 0.173%

action words relating to investigations or explorations

technical terms or descriptors related to integration and participation

No Comments

No Known Activations