INDEX

Explanations

posts or articles and their details

oai_token-act-pair · gpt-3.5-turbo

descriptive phrases about the status and content of articles or posts

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

 chimpanzees

-0.72

TVs

-0.72

unts

-0.67

aterasu

-0.66

 Macy

-0.63

 appliances

-0.62

 machines

-0.61

 cabinets

-0.60

 virginity

-0.60

metics

-0.59

POSITIVE LOGITS

 Blog

0.90

 Column

0.90

 reprinted

0.87

 Editorial

0.87

 blog

0.85

 titled

0.84

 authored

0.79

 excerpt

0.77

odox

0.76

 reprint

0.76

Activations Density 0.347%

posts or articles and their details

descriptive phrases about the status and content of articles or posts

No Comments

No Known Activations