INDEX

Explanations

code snippets or programming syntax patterns

oai_token-act-pair · gpt-3.5-turbo

occurrences of special symbols and pseudocode elements

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

 ILCS

-0.68

heny

-0.67

hips

-0.67

grading

-0.66

atform

-0.66

 relations

-0.66

asses

-0.64

actionDate

-0.62

Blocks

-0.62

 necks

-0.61

POSITIVE LOGITS

ãĥĥãĥĪ

0.79

repeat

0.74

attribute

0.70

ãĤ¦ãĤ¹

0.70

MET

0.70

 Schwarz

0.69

Miy

0.67

 Canaver

0.66

ãĥīãĥ©

0.66

TEXT

0.65

Activations Density 0.125%

code snippets or programming syntax patterns

occurrences of special symbols and pseudocode elements

No Comments

No Known Activations