INDEX

Explanations

words related to actions or attempts against a particular entity or issue

oai_token-act-pair · gpt-3.5-turbo

references to an unspecified subject related to problems or issues that need addressing

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Aluminum

-0.70

 Spoon

-0.70

 Nano

-0.69

 Pirates

-0.66

holding

-0.64

 Firefly

-0.64

 Sunshine

-0.64

 Kang

-0.62

tones

-0.62

adelphia

-0.61

POSITIVE LOGITS

self

0.99

 oneself

0.98

alian

0.85

atic

0.82

 ourselves

0.80

 accordingly

0.80

atical

0.76

 anew

0.76

 altogether

0.75

 firsthand

0.74

Activations Density 0.198%

words related to actions or attempts against a particular entity or issue

references to an unspecified subject related to problems or issues that need addressing

No Comments

No Known Activations