INDEX

Explanations

phrases related to unexpected outcomes or transformations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "turn" and its variations, indicating changes or transformations in states or situations

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

è¦ļéĨĴ

-1.03

heed

-0.67

icipated

-0.64

merga

-0.64

ording

-0.64

Ys

-0.61

pton

-0.61

ority

-0.60

cussion

-0.60

erella

-0.60

POSITIVE LOGITS

 into

1.09

 sour

1.09

 ugly

0.99

 upside

0.95

 violent

0.86

 inward

0.85

 INTO

0.84

coat

0.82

 stale

0.82

 heads

0.79

Activations Density 0.032%

phrases related to unexpected outcomes or transformations

instances of the word "turn" and its variations, indicating changes or transformations in states or situations

No Comments

No Known Activations