INDEX

Explanations

academic and professional titles and affiliations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

words related to affiliations or institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 overnight

-0.69

 alone

-0.67

 forever

-0.67

 drunk

-0.65

 grab

-0.65

 rails

-0.64

 normally

-0.64

 reverted

-0.64

 ransom

-0.63

 dwarves

-0.63

POSITIVE LOGITS

Science

1.11

Bio

1.08

negie

1.05

Research

1.04

Tech

1.04

Study

1.03

heastern

1.02

University

1.01

 Faculty

0.99

ournals

0.95

Activations Density 0.352%

academic and professional titles and affiliations

words related to affiliations or institutions

No Comments

No Known Activations