INDEX

Explanations

words related to news articles and headlines

oai_token-act-pair · gpt-3.5-turbo

abbreviations or identifiers related to location or organization names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

stood

-0.74

 Bulg

-0.67

wagen

-0.67

felt

-0.63

ttes

-0.61

stay

-0.60

except

-0.60

opausal

-0.59

Redditor

-0.58

auts

-0.58

POSITIVE LOGITS

BRE

1.06

CLAIM

1.04

COL

1.03

HAM

1.01

OF

0.99

FER

0.98

AM

0.96

VER

0.96

BR

0.94

AN

0.94

Activations Density 0.099%

words related to news articles and headlines

abbreviations or identifiers related to location or organization names

No Comments

No Known Activations