INDEX

Explanations

names of prominent individuals and organizations, particularly ones with historical or political significance

oai_token-act-pair · gpt-3.5-turbo

names of prominent individuals or families associated with wealth and influence

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

 tant

-0.70

 Hasan

-0.68

 subp

-0.68

Lago

-0.67

 Plato

-0.67

 ÏĦ

-0.63

most

-0.62

uminium

-0.60

 Calais

-0.59

 Kashmir

-0.59

POSITIVE LOGITS

kefeller

1.18

enthal

1.13

helle

1.04

otine

0.86

kef

0.85

hester

0.83

 tremend

0.82

hovah

0.81

berman

0.80

ental

0.79

Activations Density 0.027%

names of prominent individuals and organizations, particularly ones with historical or political significance

names of prominent individuals or families associated with wealth and influence

No Comments

No Known Activations