INDEX

Explanations

words related to complaints or complications

oai_token-act-pair · gpt-3.5-turbo

terms related to complaints or issues raised

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

EStream

-0.82

lein

-0.74

selection

-0.72

WARE

-0.69

plane

-0.68

 garlic

-0.67

»Ĵ

-0.67

shock

-0.67

BALL

-0.66

©¶æ

-0.66

POSITIVE LOGITS

icated

1.09

aint

1.09

ications

1.06

imentary

1.04

aints

1.03

iments

1.00

iment

0.98

icit

0.95

ained

0.93

icating

0.92

Activations Density 0.018%

words related to complaints or complications

terms related to complaints or issues raised

No Comments

No Known Activations