INDEX

Explanations

politically-oriented words related to different ideologies

oai_token-act-pair · gpt-3.5-turbo

references to various political and ideological groups

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

Delivery

-0.79

Owner

-0.74

STA

-0.65

INAL

-0.64

wise

-0.62

 geological

-0.61

inventoryQuantity

-0.61

 circumstances

-0.60

CHECK

-0.60

 Bulldogs

-0.60

POSITIVE LOGITS

ervatives

1.44

ervative

1.39

aurus

1.28

paces

1.25

pace

0.98

 rejoice

0.97

hip

0.96

chool

0.96

mith

0.96

0.94

Activations Density 0.098%

politically-oriented words related to different ideologies

references to various political and ideological groups

No Comments

No Known Activations