INDEX

Explanations

words related to technology or references to products or services

oai_token-act-pair · gpt-3.5-turbo

terms related to unique identifiers or titles of works

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

 carbohyd

-0.79

 volunte

-0.73

 substantive

-0.73

ailability

-0.73

 nursing

-0.72

udeb

-0.72

umption

-0.71

 customary

-0.70

 serious

-0.69

lying

-0.69

POSITIVE LOGITS

 Reborn

1.04

 Worlds

0.99

 Valley

0.99

 Imaging

0.98

Tree

0.97

 Records

0.94

 Junction

0.93

 Strikes

0.93

 Labs

0.93

 Warfare

0.93

Activations Density 0.219%

words related to technology or references to products or services

terms related to unique identifiers or titles of works

No Comments

No Known Activations