INDEX

Explanations

words related to foundational aspects, origins, or primary causes

oai_token-act-pair · gpt-3.5-turbo

references to foundational or fundamental concepts, often related to causes or origins

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

eers

-0.73

psey

-0.72

hemat

-0.70

 disadvant

-0.69

 Tonight

-0.67

 Dragonbound

-0.66

glers

-0.65

abwe

-0.65

ammy

-0.65

ques

-0.64

POSITIVE LOGITS

kit

1.03

 canal

1.00

beer

0.96

stock

0.95

 cellar

0.92

stocks

0.90

 Canal

0.81

arious

0.81

 Cause

0.80

less

0.78

Activations Density 0.040%

words related to foundational aspects, origins, or primary causes

references to foundational or fundamental concepts, often related to causes or origins

No Comments

No Known Activations