INDEX

Explanations

words related to technical or specific language, possibly related to programming or technical manuals

oai_token-act-pair · gpt-3.5-turbo

words related to automatic processes or systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

cat

-0.69

str

-0.68

 rough

-0.61

 scheme

-0.61

 cruel

-0.60

 scor

-0.60

 tort

-0.59

 torn

-0.59

Dh

-0.59

 frig

-0.58

POSITIVE LOGITS

omatic

4.70

oms

1.33

omal

1.15

om

1.14

omial

1.14

matic

1.13

oma

1.13

automatic

1.13

onomous

1.11

onic

1.08

Activations Density 0.010%

words related to technical or specific language, possibly related to programming or technical manuals

words related to automatic processes or systems

No Comments

No Known Activations