INDEX

Explanations

phrases related to individuals named "Pete."

oai_token-act-pair · gpt-3.5-turbo

words or names related to competitive contexts

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

 point

-0.67

 relief

-0.64

hus

-0.63

 overs

-0.61

 custody

-0.60

 points

-0.60

 Lowry

-0.59

 mills

-0.59

 Magn

-0.58

 caution

-0.57

POSITIVE LOGITS

ete

4.57

eta

1.61

et

1.59

eting

1.57

etry

1.55

ets

1.52

eto

1.50

eters

1.38

eme

1.33

ethe

1.30

Activations Density 0.005%

phrases related to individuals named "Pete."

words or names related to competitive contexts

No Comments

No Known Activations