INDEX

Explanations

words related to a specific focus or orientation within a context

oai_token-act-pair · gpt-3.5-turbo

descriptors that indicate a focus or orientation towards specific themes or characteristics

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

ricanes

-0.76

ãģį

-0.76

lost

-0.75

cell

-0.74

inus

-0.71

atel

-0.70

osponsors

-0.69

kar

-0.68

dust

-0.68

ailable

-0.67

POSITIVE LOGITS

 approach

0.99

 mindset

0.90

 Approach

0.89

 environment

0.85

 policing

0.83

 nature

0.82

 atmosphere

0.82

 attitude

0.80

 worldview

0.80

 upbringing

0.78

Activations Density 0.091%

words related to a specific focus or orientation within a context

descriptors that indicate a focus or orientation towards specific themes or characteristics

No Comments

No Known Activations