INDEX

Explanations

mentions of statements or reports by authorities

oai_token-act-pair · gpt-3.5-turbo

statements made by authorities or experts

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

Finish

-0.76

 sank

-0.75

imeter

-0.74

walk

-0.74

 sinks

-0.72

Chuck

-0.62

enders

-0.61

anyahu

-0.60

nery

-0.60

Draft

-0.59

POSITIVE LOGITS

 Computing

0.71

 IEEE

0.70

oret

0.69

 acron

0.67

lict

0.62

ynes

0.62

 Liter

0.61

 Razor

0.61

ernel

0.60

ript

0.60

Activations Density 0.000%

mentions of statements or reports by authorities

statements made by authorities or experts

No Comments

No Known Activations