INDEX

Explanations

words related to questioning the truth or validity of statements or claims

oai_token-act-pair · gpt-3.5-turbo

the word "that."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hens

-0.72

ãĥ¼ãĤ¯

-0.71

tails

-0.71

estern

-0.68

atur

-0.67

uty

-0.67

oufl

-0.67

oses

-0.65

oran

-0.64

mes

-0.64

POSITIVE LOGITS

 they

0.87

 there

0.82

 pesky

0.79

 fateful

0.78

 although

0.76

genre

0.74

 THEY

0.73

 someday

0.73

soever

0.72

 elusive

0.71

Activations Density 0.276%

words related to questioning the truth or validity of statements or claims

the word "that."

No Comments

No Known Activations