INDEX

Explanations

negative or problematic issues

oai_token-act-pair · gpt-3.5-turbo

phrases highlighting problems or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ertodd

-0.69

roxy

-0.69

eday

-0.68

 Alive

-0.67

steen

-0.66

ULTS

-0.64

chedel

-0.61

 Alam

-0.60

ilver

-0.60

stocks

-0.60

POSITIVE LOGITS

 drawback

1.82

 problem

1.75

 disadvantage

1.70

 limitation

1.59

 issue

1.51

 drawbacks

1.51

 disadvantages

1.49

 hurdle

1.46

 problems

1.45

 inconvenience

1.44

Activations Density 0.455%

negative or problematic issues

phrases highlighting problems or issues

No Comments

No Known Activations