INDEX

Explanations

phrases used to express a common theme or idea across different contexts

oai_token-act-pair · gpt-3.5-turbo

phrases that convey common expressions or sayings about predictions and consequences

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

ullah

-0.75

 Horus

-0.70

ament

-0.70

icio

-0.69

ricted

-0.65

ussen

-0.65

ature

-0.64

urated

-0.62

ificent

-0.62

ricting

-0.61

POSITIVE LOGITS

Ń·

0.95

verning

0.93

lems

0.92

 overboard

0.92

vt

0.91

©¶æ

0.78

 Forth

0.78

 forth

0.78

OHN

0.76

ggle

0.76

Activations Density 0.089%

phrases used to express a common theme or idea across different contexts

phrases that convey common expressions or sayings about predictions and consequences

No Comments

No Known Activations