INDEX

Explanations

the word "Hin" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĤº

-0.68

Reference

-0.66

urally

-0.62

Loading

-0.62

 Anarchy

-0.62

 manifest

-0.62

sur

-0.62

Balance

-0.60

pair

-0.60

multipl

-0.60

POSITIVE LOGITS

hin

1.21

osaurs

1.12

eman

1.06

emen

0.98

ney

0.89

agar

0.87

hoe

0.84

estones

0.84

ners

0.82

emi

0.81

Activations Density 0.008%

the word "Hin" in various contexts

references to specific individuals or names

No Comments

No Known Activations