INDEX

Explanations

text related to photos or images

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Photo" and associated fragments or notes, possibly indicating a focus on visual elements or imagery in the text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hement

-0.77

ngth

-0.58

hower

-0.57

 Hollow

-0.56

full

-0.55

oway

-0.55

anded

-0.54

worth

-0.54

 fronts

-0.53

 Lean

-0.53

POSITIVE LOGITS

Phys

0.67

Associated

0.62

Charles

0.62

 Continue

0.61

VERTIS

0.60

 guiActiveUnfocused

0.60

window

0.60

 GOODMAN

0.60

 VIDEOS

0.60

eway

0.59

Activations Density 0.021%

text related to photos or images

instances of the word "Photo" and associated fragments or notes, possibly indicating a focus on visual elements or imagery in the text

No Comments

No Known Activations