INDEX

Explanations

phrases related to government entities, specifically the State Department

oai_token-act-pair · gpt-3.5-turbo

references to the State Department

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

amaz

-0.90

juries

-0.79

âĸ¬âĸ¬

-0.71

 Wong

-0.71

é»Ĵ

-0.68

ãĤ©

-0.68

 Packs

-0.67

teenth

-0.67

à¸

-0.67

AMD

-0.65

POSITIVE LOGITS

 Correspond

0.92

 diplomats

0.85

 secretaries

0.82

 envoy

0.81

 bureaucracy

0.81

 spokeswoman

0.81

 embassies

0.80

 spokesperson

0.80

 staffer

0.79

 embassy

0.79

Activations Density 0.041%

phrases related to government entities, specifically the State Department

references to the State Department

No Comments

No Known Activations