INDEX

Explanations

location-related information, such as places, addresses, and descriptions

oai_token-act-pair · gpt-3.5-turbo

references to locations and spaces related to community or social activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Gaddafi

-0.73

 Conservatives

-0.67

 Marxism

-0.64

democracy

-0.63

 Liberals

-0.63

ihadi

-0.62

 Britons

-0.62

 Luxem

-0.62

Brexit

-0.61

 purported

-0.61

POSITIVE LOGITS

 refill

0.83

 snack

0.78

 snacks

0.77

aturdays

0.72

 handy

0.71

 chores

0.71

 babys

0.70

 veggies

0.70

 horizont

0.69

 meals

0.69

Activations Density 1.559%

location-related information, such as places, addresses, and descriptions

references to locations and spaces related to community or social activities

No Comments

No Known Activations