INDEX

Explanations

phrases related to specific entities or organizations

oai_token-act-pair · gpt-3.5-turbo

references indicating groups or types of entities

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

zzle

-0.79

iness

-0.76

hood

-0.72

opus

-0.69

pless

-0.69

Ã«

-0.68

erity

-0.67

nah

-0.67

formance

-0.65

rogram

-0.63

POSITIVE LOGITS

 kinds

0.87

 findings

0.83

 fellows

0.81

 results

0.79

 sorts

0.78

 matters

0.77

 facts

0.76

 truths

0.75

 endeavors

0.74

 developments

0.74

Activations Density 0.073%

phrases related to specific entities or organizations

references indicating groups or types of entities

No Comments

No Known Activations