INDEX

Explanations

names of professions and specific persons

oai_token-act-pair · gpt-3.5-turbo

references to specific professionals or expert roles in various contexts

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

needed

-0.60

awa

-0.59

anwhile

-0.56

ruby

-0.54

 Unloaded

-0.53

oother

-0.53

amera

-0.50

jad

-0.49

 warming

-0.48

valued

-0.48

POSITIVE LOGITS

 consortium

0.63

 standpoint

0.59

 manufacturer

0.54

 contractor

0.53

icus

0.52

 landowners

0.51

 margin

0.50

 eminent

0.50

alth

0.48

external

0.48

Activations Density 1.365%

names of professions and specific persons

references to specific professionals or expert roles in various contexts

No Comments

No Known Activations