INDEX

Explanations

references to a male individual with strong positive associations, possibly related to admiration, support, or recognition

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual

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

services

-0.71

odge

-0.66

Services

-0.62

give

-0.61

formed

-0.61

 Carrie

-0.61

ibrary

-0.60

icion

-0.60

AWS

-0.60

mble

-0.60

POSITIVE LOGITS

 personally

0.86

atic

0.84

atically

0.83

 Majesty

0.83

panic

0.81

 orally

0.73

ading

0.73

atics

0.72

alian

0.71

 tremend

0.71

Activations Density 0.082%

references to a male individual with strong positive associations, possibly related to admiration, support, or recognition

references to a specific individual

No Comments

No Known Activations