INDEX

Explanations

words ending in "ly" that describe something

oai_token-act-pair · gpt-3.5-turbo

adverbs that modify descriptions or actions

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

apixel

-0.76

venge

-0.68

 paycheck

-0.67

rouse

-0.62

utm

-0.61

ocene

-0.60

 throats

-0.60

elf

-0.59

igers

-0.59

 cavity

-0.59

POSITIVE LOGITS

 importantly

0.77

Ħ¢

0.74

 interestingly

0.71

 controvers

0.67

 though

0.67

 Ironically

0.67

NESS

0.67

ional

0.66

entimes

0.65

nown

0.64

Activations Density 0.050%

words ending in "ly" that describe something

adverbs that modify descriptions or actions

No Comments

No Known Activations