INDEX

Explanations

statements related to truthfulness

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "truth."

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

uled

-0.86

wana

-0.77

capacity

-0.75

avy

-0.73

 Rivals

-0.70

jin

-0.67

urations

-0.67

joining

-0.66

ATA

-0.66

oyal

-0.65

POSITIVE LOGITS

fulness

1.24

fully

1.06

psons

0.92

lessly

0.80

 seeker

0.79

 truth

0.78

ulent

0.78

 serum

0.76

iness

0.76

telling

0.75

Activations Density 0.014%

statements related to truthfulness

occurrences of the word "truth."

No Comments

No Known Activations