INDEX

Explanations

references to reports or studies released by specific organizations

oai_token-act-pair · gpt-3.5-turbo

references to reports and their sources

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

ados

-0.75

 rever

-0.73

 whilst

-0.70

 alike

-0.70

qus

-0.69

 namely

-0.69

 engulf

-0.67

nuts

-0.67

rats

-0.66

aji

-0.66

POSITIVE LOGITS

 National

1.15

 Department

1.14

 Ministry

1.09

 Bureau

1.09

 Association

1.05

 Legislative

1.04

 Associated

1.03

 ACLU

1.01

 Electoral

1.01

 Office

1.01

Activations Density 0.439%

references to reports or studies released by specific organizations

references to reports and their sources

No Comments

No Known Activations