INDEX

Explanations

words related to justification or exoneration

oai_token-act-pair · gpt-3.5-turbo

terms related to validation and vindication

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

atomic

-0.71

pmwiki

-0.70

BOOK

-0.70

duino

-0.69

bnb

-0.69

sites

-0.67

cock

-0.67

abwe

-0.66

ascus

-0.65

 athlet

-0.63

POSITIVE LOGITS

ication

1.16

ictive

1.11

icated

1.04

icators

0.96

ict

0.95

icate

0.94

icator

0.93

ications

0.90

iary

0.90

ECT

0.89

Activations Density 0.064%

words related to justification or exoneration

terms related to validation and vindication

No Comments

No Known Activations