INDEX

Explanations

words and phrases related to thoughts or considerations

oai_token-act-pair · gpt-3.5-turbo

thoughts and perceptions related to possibilities and sights

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

vic

-0.76

arta

-0.67

ufact

-0.67

ilver

-0.66

ummies

-0.65

este

-0.65

visory

-0.65

oche

-0.64

imon

-0.64

hops

-0.64

POSITIVE LOGITS

 itself

0.85

enance

0.73

 afforded

0.71

lessness

0.69

 elevated

0.68

 implication

0.67

ossibility

0.66

IENCE

0.66

 posed

0.65

unn

0.65

Activations Density 0.183%

words and phrases related to thoughts or considerations

thoughts and perceptions related to possibilities and sights

No Comments

No Known Activations