INDEX

Explanations

mentions of experts offering insights or opinions

oai_token-act-pair · gpt-3.5-turbo

references to expert opinions or assessments

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

Interstitial

-0.88

HAHA

-0.69

TO

-0.68

gress

-0.67

nant

-0.66

Ore

-0.64

Neh

-0.63

HAHAHAHA

-0.62

ãĥª

-0.62

Complete

-0.62

POSITIVE LOGITS

 speculate

1.17

 alike

1.08

paces

1.00

 differ

0.95

hips

0.94

 contend

0.91

 consulted

0.91

 argue

0.90

 anticipate

0.90

 believe

0.89

Activations Density 0.154%

mentions of experts offering insights or opinions

references to expert opinions or assessments

No Comments

No Known Activations