INDEX

Explanations

statements that attribute information to a specific source or author

oai_token-act-pair · gpt-3.5-turbo

sources of information or references to claims

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

-0.78

 realise

-0.69

ãĥĵ

-0.69

ãĥĨãĤ£

-0.67

bers

-0.64

bered

-0.63

 ashore

-0.62

ãģĹ

-0.61

blast

-0.60

POSITIVE LOGITS

chwitz

0.98

ety

0.84

tesy

0.80

Style

0.80

uilt

0.76

achev

0.75

ymm

0.75

rary

0.75

hement

0.75

iosity

0.75

Activations Density 0.024%

statements that attribute information to a specific source or author

sources of information or references to claims

No Comments

No Known Activations