INDEX

Explanations

locations, especially related to political events

oai_token-act-pair · gpt-3.5-turbo

state names and their corresponding political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Britann

-0.70

Japan

-0.67

 Tokyo

-0.66

 Hiroshima

-0.66

 decomp

-0.65

Tok

-0.65

ital

-0.64

ammers

-0.64

Toronto

-0.63

 Manila

-0.63

POSITIVE LOGITS

 senator

1.27

 Republican

1.24

 elector

1.19

 Democratic

1.19

 Senator

1.19

 caucuses

1.19

 gubernatorial

1.18

 congressman

1.18

 caucus

1.16

 governor

1.15

Activations Density 0.108%

locations, especially related to political events

state names and their corresponding political context

No Comments

No Known Activations