INDEX

Explanations

names of individuals associated with various topics or events

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals in various contexts

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

 Rivals

-0.53

 opio

-0.52

 dwar

-0.49

 whirlwind

-0.49

ichita

-0.49

 pestic

-0.48

 denomin

-0.48

 cryst

-0.47

 sugg

-0.47

VIDIA

-0.46

POSITIVE LOGITS

 goodbye

0.68

 differently

0.62

 squarely

0.61

ById

0.59

 hostage

0.59

 correctly

0.58

 onto

0.57

selves

0.57

 badge

0.57

 into

0.55

Activations Density 1.117%

names of individuals associated with various topics or events

mentions of specific individuals in various contexts

No Comments

No Known Activations